Coin Toss: Game-Theoretic Fairness and Privacy.

PASSCODE: 985224

Abstract: Secure coin-toss protocols are typically modeled as input-less functionalities where parties with no secret inputs want to toss a fair coin collaboratively. In the presence of a dishonest majority, it is well-known that a strongly fair coin-tossing protocol is impossible. To circumvent this negative result, recent works have explored a relaxed notion of game-theoretic fairness, where adversaries are modeled as rational parties with preferences for specific outcomes, seeking to bias the coin in their favor. These preferences, which can encode secret information, challenge the conventional view of coin-tossing as an input-less functionality.

We initiate a comprehensive study of privacy-preserving game-theoretic fairness for multi-party multi-sided coin toss. In this talk, I will present the landscape of privacy-preserving game-theoretically fair coin-toss protocols against dishonest majority. Specifically, I will show how to build game-theoretically fair coin-toss protocols where everyone’s preference is public, and then show how to upgrade the protocol to achieve privacy while maintaining the same corruption resilience.

Based on joint work with Pedro Branco (Bocconi University), Pratik Soni (UUtah), Sri AravindaKrishnan Thyagarajan (USydney).